AI in Recommendation Systems Quiz

Answers and explanations appear as you take the quiz.

1. What is collaborative filtering based on?
   • Patterns of preference across many users
   • The colors used in an item's image
   • The length of product descriptions
   • Randomly shuffling the catalog
2. What does content-based filtering use to make recommendations?
   • Only the time of day
   • Features of the items and your past preferences
   • The stock price of the company
   • A coin flip
3. What problem does matrix factorization solve in recommenders?
   • Encrypting user passwords
   • Compressing video files
   • Turning a sparse user-item matrix into compact latent vectors
   • Routing internet traffic
4. What is the 'cold start' problem?
   • Servers being too cold to run
   • A virus that freezes accounts
   • Recommendations only working in winter
   • Difficulty recommending for brand-new users or items with no history
5. What competition helped popularize recommendation research?
   • The Netflix Prize
   • The Turing Award
   • The Nobel Prize
   • The Eurovision contest
6. What did Meyer and colleagues propose the Average Measure of Impact to evaluate?
   • Image classification accuracy
   • The impact of personalized recommendations on usefulness and trust
   • The number of ratings stored in a matrix
   • The GPU memory used by a recommender
7. Which metric is designed to assess the quality of a ranked top-K recommendation list?
   • Mean absolute error on a single rating only
   • Pixel accuracy
   • NDCG@K
   • Word error rate
8. Why can a low rating-prediction error fail to guarantee a good recommendation list?
   • RMSE directly measures user satisfaction for every ranking
   • Every user sees every catalog item
   • A rating model cannot use user-item data
   • Rating accuracy and top-item ranking quality are different objectives